Conventionally, there has been known a method of estimating a full-charge capacity of a secondary battery based on a difference of SOC between two points and a current integration value between the two points (see JP 2003-68369 A). This estimation method assumes that there is a region in which an OCV-SOC table does not change due to capacity degradation.